Combining Bayesian Beliefs and Willingness to Bet to Analyze Attitudes towards Uncertainty

Speaker: Mohammed ABDELLAOUI
CNRS & HEC, Paris

9 November 2007

Many deviations from Bayesianism have been found for choices under uncertainty with unknown probabilities (“ambiguity”). General choice-based models and qualitative empirical tests have been provided. This paper introduces an operational procedure for quantitatively measuring attitudes towards uncertainty and ambiguity, allowing for exact predictions. We first identify sources of uncertainty, comprising events generated by a common mechanism and with a uniform degree of ambiguity. For such sources we can define choice-based probabilities. Key in our approach is the introduction of source functions, converting choice-based probabilities into decision weights. Using these source functions, attitudes towards uncertainty and ambiguity can be conveniently analyzed and displayed in tractable graphs. We demonstrate the implementability of our approach in an experiment. The prevailing phenomena comprise both aversion and insensitivity towards ambiguity.

Dépendance de l'utilité par rapport aux états : questions et enjeux dans la tradition savagienne

Speaker: Brian HILL

12 October 2007 - From 16h15 to 18h15